|Publication number||US7330467 B2|
|Application number||US 10/396,433|
|Publication date||Feb 12, 2008|
|Filing date||Mar 26, 2003|
|Priority date||Mar 26, 2003|
|Also published as||US20040190502|
|Publication number||10396433, 396433, US 7330467 B2, US 7330467B2, US-B2-7330467, US7330467 B2, US7330467B2|
|Original Assignee||Altera Corporation|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (11), Referenced by (21), Classifications (10), Legal Events (6)|
|External Links: USPTO, USPTO Assignment, Espacenet|
The invention relates generally to the field of network communications. More specifically, the invention relates to a system and method for switching communications traffic at a node in a network.
Techniques for switching communications traffic at a node in a network are known. As illustrated in
The primary switch fabric card 125 and back-up switch fabric card 130 are configured to redirect network traffic (data) to one or more line cards. In turn, the line cards transmit the network traffic to a next or final destination node on the network. The primary switch fabric card 125 and back-up switch fabric card 130 can include, for example, crossbars or shared memory devices. The line cards 105, 110, 115, 120, 135, and 140 are used for buffering network traffic on either side of the switch fabric, and for performing other functions. Line cards typically serve both ingress and egress functions (i.e., for incoming and outgoing traffic, respectively).
Router communications can be separated into three categories: management plane control plane and data plane communications. The management plane is an interface between management functions external to the router (e.g., network servers or clients) and the router controller(s) for management of the router. For example, chassis configuration parameters that are derived from Service Level Agreements (SLA's) are communicated to the router on the management plane. The control plane uses local signaling protocols or other messages to control the resources of a router in accordance with the specified configuration. The data plane carries network data that is being redirected (or forwarded) by the router to the next or final destination node in the network.
In the data plane, network traffic is received at a line card, processed through the primary switch fabric card 125 (or back-up switch fabric card 130 if the primary switch fabric card 125 is not functioning), and forwarded to the (same or different) line card. The controller card 145 typically hosts switching application protocols (e.g., Open Shortest Path First (OSPF), Routing Information Protocol (RIP), Multi-Protocol Label Switching (MPLS) or other protocols) and generates messages to the line cards 105, 110, 115, 120, 135 and 140, the primary switch fabric card 125 and the back-up switch fabric card 130 in the control plane.
Known systems and methods for switching communications traffic have various disadvantageous. For example, in the control plane of known routers, a switching application must establish communications with each of the device drivers on line cards 105, 110, 115, 120, 135 and 140, the primary switch fabric card 125 and the back-up switch fabric card 130. Such a control scheme adds complexity to the development of switching applications. For example, where a line card becomes non-functional or is removed from the router chassis, the switching application must first identify the non-functional or removed card, then notify each of the device drivers associated with cards in chassis 150. Likewise, when a new card is added to chassis 150, the switching application must execute a lengthy process of registration, initialization, and configuration involving each of the device drivers. Including such complexities may extend the time-to-market for new switching applications under development.
Therefore, a need exists for a system and method to simplify the interface between a switching application and device drivers in the control plane of a network switch or router.
One embodiment of the invention provides a proxy driver in the control plane of the router or switch providing a centralized Application Program Interface (API) between a switching application and multiple device drivers in a router chassis. Another embodiment of the invention provides a centralized method for enforcing a topology in the control plane of a router. Another embodiment of the invention provides a centralized method in the control plane for the discovery of resources, for example as cards are added or removed from a router chassis.
The features and advantages of the invention will become apparent from the following drawings and detailed description.
The invention is directed to improving the efficiency of a router, switch, or similar network element (NE) configured to switch data in a communications network. The terms router, switch, and NE are used herein interchangeably.
Embodiments of the invention relate to the control plane of a router. In one embodiment of the invention, a proxy driver in the controller enables a centralized API to distribute resources within the chassis. Such a scheme advantageously eliminates the need for the application program to make local software calls to each line card and/or switch fabric card. In contrast to other proxy driver schemes, some embodiments of the invention use a single proxy driver to control multiple resource drivers within the router. In typical embodiments, the single proxy driver is located on a controller card, and the multiple resource drivers are on multiple line and/or switch fabric cards. Moreover, each line card or switch fabric card can include multiple resource drivers. The single proxy driver thus simplifies the application environment for intelligent switch fabric or other router products.
Each of the router resources (e.g., line cards, switch cards, and controllers) include middleware components in communication with the proxy driver. Together, the proxy driver and middleware components enable enforcement of a topology within the router chassis and facilitate intelligent discovery of chassis resources.
Subheadings used in this section are for organizational convenience, and are not meant to indicate that the disclosure of any particular feature is limited to any particular subheading used herein.
In operation, the EMS 230 converts high-level policy and service level agreements (SLA's) into chassis configuration requirements, and communicates those requirements to the primary controller card 220 and/or the back-up controller card 215. Then, under the control of the primary controller card 220 or the back-up controller card 215, network traffic is switched to and/or from line card 205 through the primary switch fabric card 210 or the back-up switch fabric card 225.
Line card 205 is an ingress line card, an egress line card, or (more typically) an ingress/egress line card. In the latter case, traffic is received in one direction and sent in the opposite direction. Line card 205 includes fabric interface device (FID) driver 275, FID chip 260, and external communication ports (not shown). The line card 205 receives or sends traffic over the external communication ports and, for example, buffers incoming and/or outgoing traffic in the node and schedules data through the primary switch fabric card 210 or back-up switch fabric card 225.
FID driver 275 and FID chip 260 control the transfer of packets, cells, or other groupings of data to and/or from the primary switch fabric card 210 or back-up switch fabric card 225.
The external communication ports (not shown) provide an interface between data communication networks and the router 200. The external communication ports (not shown) can be, for example, compatible with GigE, 10/100 Ethernet, TI, E1, or other communication protocols.
The primary switch fabric card 210 includes switch chip driver 280 and switch chip 265; back-up switch fabric card 225 includes switch chip driver 295 and switch chip 270. Switch chips 265 and 270 can be, for example, crossbars for switching data packets or other divisions of data between input ports (not shown) and output ports (not shown) of switch chips 265 and 270. Switch chip drivers 280 and 295 are software or other functional modules configured to control the switch chips 265 and 270, respectively. Back-up switch fabric card 225 is a redundant capability to be used when primary switch fabric card 210 is removed from the router chassis or is otherwise non-functional.
Primary controller card 220 and back-up controller card 215 include proxy drivers 290 and 285, respectively. Proxy drivers 290 and 285 provide a centralized API in primary controller card 220 and back-up controller card 215, respectively, and may include a master topology definition for the control plane of the router as described with reference to
Line card 205, switch fabric card 210, back-up switch fabric card 225, primary controller card 220 and back-up controller card 215 each include middleware components 235, 240, 255, 250, and 245, respectively. Middleware components 235, 240, 255, 250, and 245 enable the proxy driver 290 or 285 to communicate with the FID driver 275, switch chip driver 280 and switch chip driver 295. Thus, the middleware components provide an intelligent conduit between the proxy drivers, the FID drivers, and the switch chip drivers in the router.
Alternative embodiments to that illustrated in
In embodiments of the invention, the proxy drivers 285 and 290, and middleware components 235, 240, 245, 250, and 255 enforce topology rules and perform intelligent discovery as discussed below with reference to
As indicated above, proxy drivers 285 and 290 provide a centralized API in a router. In one embodiment of the invention, proxy drivers 285 and 290 dictate a messaging topology between cards in the control plane 240 of a router chassis. A topology is implemented by session objects that provide links between middleware objects in the cards of a router chassis.
The star topology illustrated by flow diagram 340 requires fewer sessions to achieve the same routing characteristics of a full mesh topology. For example, the topology illustrated by flow diagram 340 includes four links, whereas the topology illustrated by flow diagram 335 includes ten links. To pass most messages in a star topology, however, data packets or cells traverse at least two hops. For example, a message between back-up line card 325 and primary line card 315 first hops from back-up line card 325 to primary CPU 305, then hops from primary CPU 305 to primary line card 315. In many cases, a star topology can gracefully degrade if an individual resource is lost. If a card is lost at the center of the star topology, however, then no further cards in chassis 330 are reachable. For example, in
Hybrid topologies such as the one illustrated in flow diagram 345 are also possible. Note, however, that the hybrid messaging topology illustrated in flow diagram 345 typically involve intermediate hops for many different messages between cards, and that a failure at slot 3 in the illustrated example would be fatal.
In one embodiment of the invention, a system designer or other user can specify any of the topologies represented by flow diagrams 335, 340 and 345. Moreover, the centralized control provided by the proxy driver advantageously supports an environment where messaging topologies other than those described above can be readily implemented. Although not shown in
Activation causes the line cards and switch fabric cards to transmit a refresh message to the controller card (not shown in
Accordingly, in a chassis with an operational controller, the controller periodically transmits the topology definition (or a sub-set thereof) to the line cards and switch fabric cards in response to polling from the line cards and switch fabric cards.
In operation, a topology is defined in the master topology definition of proxy driver 520. In the illustrated embodiment, the master topology definition of proxy driver 520 defines sessions between: logical cards 1 and 3 (1-3); logical cards 2 and 1 (2-1); logical cards 2 and 3 (2-3); logical cards 2 and 4 (2-4); and logical cards 3 and 4 (3-4). In the illustrated embodiment, the master topology definition also includes parameters to indicate active (A) or inactive (I) sessions. Because logical card 4 is not installed, sessions related to the uninstalled card 540 are inactive (I).
Middleware components 510, 525, and 535 instantiate the chassis topology in each of the cards 505, 515, and 530, respectively. In the illustrated embodiment, only relevant portions of the overall topology for the control plane are stored by each middleware component as local routing tables. Thus, from the perspective of middleware component 510, the valid sessions are: session PI between itself (logical 1) and controller card 515 (logical 2); and session P3 between itself (logical 1) and switch fabric card 530 (logical 3). Likewise, from the perspective of middleware component 535, the only valid sessions are: session P2 between itself (logical 3) and controller card 515 (logical 2); and session P3 between itself (logical 3) and line card 510 (logical 1). Middleware component 525 recognizes: session PI between itself (logical 2) and line card 505 (logical 1), and also session P2 between itself (logical 2) and switch fabric card 530 (logical 3). Thus, in the illustrated embodiment, a full mesh topology in the control plane of a router is enforced by proxy driver 520 and middleware components 510, 525, and 535.
Changes to the master topology definition of proxy driver 520 allow for alternative control plane topologies. In the illustrated embodiment of
As illustrated in
With reference to both
If it is determined in conditional step 610 that a refresh message has not been received at the controller within a predetermined time, then the process advances to conditional step 615. There are at least two ways for determining that a slot is down in step 615. In a first case, a router resource may detect a failure, and notify the controller card that the router resource is non-functional. For example, if on-card diagnostics of switch fabric card 530 detect a failed switch chip, then middleware component 535 can inform the controller card 515 that the switch fabric card 530 is non-functional. Alternatively, or in combination, the controller card 515 can poll all chassis resources at a predetermined time or interval. If a card is removed from a slot in the chassis, or if a card is no longer functioning, attached session objects will (immediately or eventually) cease to operate. Thus, if controller card 515 polled switch fabric card 530 and failed to receive the appropriate response, then the controller card 515 would have an indication the switch fabric card 530 (slot 3) is non-functional.
If it is determined in conditional step 615 that a slot-down message has been received, the process advances to step 635 where the controller card 515 determines which slot is not operational. Where a slot down message was received at the controller card 515 from another resource in the chassis, the received message may itself contain the slot information. Alternatively, where the controller discovered a non-functional card via polling, the polling logic may identify the non-functional slot.
Once it is determined which slot is not operational, the process advances to step 640 where the controller card 515 updates status parameters in the master topology definition. For example, if the controller determined that slot 3 is down, then it would change “1-3, A” to “1-3, I,” and further change “2-3, A” to “2-3, I.”
Next, the process advances to step 645 where the controller transmits a slot-down message to each middleware component previously having a valid session with the non-functional card to update local topology definitions. Thus, if it were determined that slot 3 is non-functional, the controller card 515 would notify middleware components 525 and 510. In response, middleware 525 would delete the “2-3, P2” entry from its local routing table, and middleware 510 would delete “1-3, P3” from its local routing table. In an alternative embodiment, instead of sending a slot-down message, the controller sends replacement topology definitions (or applicable sub-sets thereof) to middleware components 525 and 510.
If it is determined in step 615 that a slot-down message has not been received, the process advances to delay step 620. The process also advances to delay step 620 after executing steps 625 and 645.
Accordingly, the proxy driver and middleware components advantageously enable an intelligent discovery process enabling the router to adapt dynamically to changes in router chassis resources.
The methods described herein can be embodied in processor-executable code, and may further be stored in processor-readable medium (e.g., hard disk, CD ROM, or other storage device).
The invention described above thus overcomes the disadvantages of known systems by providing a centralized application program interface to simplify the development of switching applications and reduce time-to-market. In addition, when combined with middleware components on each card in the router chassis, topology enforcement and intelligent discovery are enabled.
While this invention has been described in various explanatory embodiments, other embodiments and variations can be effected by a person of ordinary skill in the art without departing from the scope of the invention.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US6192051||Feb 26, 1999||Feb 20, 2001||Redstone Communications, Inc.||Network router search engine using compressed tree forwarding table|
|US6314525||Oct 2, 1997||Nov 6, 2001||3Com Corporation||Means for allowing two or more network interface controller cards to appear as one card to an operating system|
|US6330599||Oct 16, 1998||Dec 11, 2001||Cisco Technology, Inc.||Virtual interfaces with dynamic binding|
|US6366587||Aug 31, 1998||Apr 2, 2002||Cisco Systems, Inc.||TDM backplane stream resource distribution system|
|US6408061||Mar 12, 1999||Jun 18, 2002||Nortel Networks Limited||Interfacing a communication switch to a non-embedded device driver via a virtual device interface|
|US6954463 *||Dec 11, 2000||Oct 11, 2005||Cisco Technology, Inc.||Distributed packet processing architecture for network access servers|
|US6956854 *||Dec 20, 2001||Oct 18, 2005||Alcatel Internetworking (Pe), Inc.||Network switching device with forwarding database tables populated based on use|
|US7010715 *||Nov 5, 2001||Mar 7, 2006||Marconi Intellectual Property (Ringfence), Inc.||Redundant control architecture for a network device|
|US7031263 *||Feb 8, 2000||Apr 18, 2006||Cisco Technology, Inc.||Method and apparatus for network management system|
|US20020132620||Mar 19, 2001||Sep 19, 2002||Ylian Saint-Hilaire||Roaming in a communication network|
|EP1077558A1||Aug 14, 2000||Feb 21, 2001||Eighty Twenty B.V.||Communication network|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US7904546||Feb 6, 2006||Mar 8, 2011||Alcatel-Lucent Usa Inc.||Managing processes on a network device|
|US8218453 *||Feb 8, 2007||Jul 10, 2012||Telefonaktiebolaget Lm Ericsson (Publ)||Network router and method of configuring a network router|
|US8560660||Dec 15, 2010||Oct 15, 2013||Juniper Networks, Inc.||Methods and apparatus for managing next hop identifiers in a distributed switch fabric system|
|US8718063||Jul 25, 2011||May 6, 2014||Juniper Networks, Inc.||Methods and apparatus related to route selection within a network|
|US8798045||Dec 29, 2008||Aug 5, 2014||Juniper Networks, Inc.||Control plane architecture for switch fabrics|
|US8918631||Mar 31, 2009||Dec 23, 2014||Juniper Networks, Inc.||Methods and apparatus for dynamic automated configuration within a control plane of a switch fabric|
|US8942559 *||May 20, 2010||Jan 27, 2015||Hewlett-Packard Development, L.P.||Switching in a network device|
|US8964733||Jul 29, 2014||Feb 24, 2015||Juniper Networks, Inc.||Control plane architecture for switch fabrics|
|US8990365||Feb 6, 2006||Mar 24, 2015||Alcatel Lucent||Processing management packets|
|US9106527||Dec 22, 2010||Aug 11, 2015||Juniper Networks, Inc.||Hierarchical resource groups for providing segregated management access to a distributed switch|
|US9240923||Dec 15, 2010||Jan 19, 2016||Juniper Networks, Inc.||Methods and apparatus for automatically provisioning resources within a distributed control plane of a switch|
|US9282060||Dec 15, 2010||Mar 8, 2016||Juniper Networks, Inc.||Methods and apparatus for dynamic resource management within a distributed control plane of a switch|
|US9344386 *||Dec 30, 2013||May 17, 2016||Wichorus, Inc.||Methods and apparatus for providing distributed load balancing of subscriber sessions in a multi-slot gateway|
|US9391796||Dec 22, 2011||Jul 12, 2016||Juniper Networks, Inc.||Methods and apparatus for using border gateway protocol (BGP) for converged fibre channel (FC) control plane|
|US20080170580 *||Jan 17, 2007||Jul 17, 2008||Goldman Stuart O||Star network control with mesh network payload delivery|
|US20100150024 *||Feb 8, 2007||Jun 17, 2010||Juttner Alpar||Network Router and Method of Configuring a Network Router|
|US20110238816 *||Dec 15, 2010||Sep 29, 2011||Juniper Networks, Inc.||Methods and apparatus for automatically provisioning resources within a distributed control plane of a switch|
|US20130266309 *||May 20, 2010||Oct 10, 2013||Nathan Binkert||Switching in a network device|
|US20140112132 *||Dec 30, 2013||Apr 24, 2014||Wichorus||Methods and Apparatus for Providing Distributed Load Balancing of Subscriber Sessions in a Multi-slot Gateway|
|CN102907057A *||May 20, 2010||Jan 30, 2013||惠普发展公司，有限责任合伙企业||Switching in a network device|
|CN102907057B *||May 20, 2010||Dec 2, 2015||惠普发展公司，有限责任合伙企业||网络装置中的交换|
|U.S. Classification||370/360, 370/392|
|International Classification||H04L12/50, H04Q11/00, G06F9/445|
|Cooperative Classification||H04Q2213/1304, H04Q2213/1305, H04Q3/54583, H04Q2213/1302|
|Mar 26, 2003||AS||Assignment|
Owner name: ZAGROS NETWORKS, INC., MARYLAND
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:SHARMA, VIPAL;REEL/FRAME:013918/0277
Effective date: 20030326
|Aug 29, 2003||AS||Assignment|
Owner name: PTS CORPORATION, CALIFORNIA
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:ZAGROS NETWORKS, INC.;REEL/FRAME:014441/0306
Effective date: 20030813
|May 12, 2004||AS||Assignment|
Owner name: PTS CORPORATION, CALIFORNIA
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:ZAGROS NETWORKS, INC.;REEL/FRAME:014623/0410
Effective date: 20030813
|May 24, 2006||AS||Assignment|
Owner name: ALTERA CORPORATION, CALIFORNIA
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:PTS CORPORATION;REEL/FRAME:017663/0320
Effective date: 20060501
|Jul 21, 2011||FPAY||Fee payment|
Year of fee payment: 4
|Jul 28, 2015||FPAY||Fee payment|
Year of fee payment: 8